Hello,
VMware-VCSA-all-8.0.2-22385739.iso
VCSA setup wizard stuck at 19% stage-2:An error occurred while starting service 'vc-ws1a-broker'
Tried modifying the deployment size from Tiny to Small, but it still didn't work.
Does anyone have insight or advice on this ?
Hello,
Lab#01: Fresh installation via VMware-VCSA-all-8.0.2-22617221.iso.
Lab#02: Update from 8.0.1 via VMware-VCSA-all-8.0.2-22617221.iso.
The problem has been solved.
Same problem bro.
Got the same issue.
Same problem.
Couple of things. @lamw talks about the new vc-ws1a-broker service on his recent blog post below.
https://williamlam.com/2023/09/what-is-vc-ws1a-broker-service-on-vcenter-server-appliance-vcsa.html
The article doesn't specifically cover the issue you're seeing but it does explain what the service is, and highlights that it does require a lot (2GB) of RAM, with VCSA 8 now requiring 14GB in total, so the first step is to ensure the ESXi host you are deploying on has sufficient RAM available to deploy the appliance.
Secondly, you may find some clues in the VCSA install log as described here. They are located at C:\Users\Username\AppData\Local\VMware\CIP\vcsaInstaller where username is the user account you used to run the installer.
HTH
Have no problem about Installing.Only Step 2:configuration went wrong.
@XZChen wrote:Have no problem about Installing.Only Step 2:configuration went wrong.
@XZChen - You should probably start a new thread, as that is a different issue.
As a sidenote, when Stage 1 completes, try resuming Stage 2 from the VAMI interface. This can fix some issues with Stage 2 not completing via the UI Installer. As above though, you should probably log a new thread.
HTH
Have the same issue and the fix for us is to remove the "search ." from the /etc/resolve.conf. It seems to get added as part of the upgrade process as it isn't present on our VCSA when running 8U1(GA/a/c).
Removing that line allows us to start the vc-ws1a-broker service and then resume the upgrade from VAMI.
Maybe I've found a better way.
During the final data migration process, check whether the vc-ws1a-broker service is in the Pending state, and check whether there is a prompt for failure to open the file in /var/log/vmware/vc-ws1a-broker/nginx-error.log. If it appears, , we can create a file as follows to make it run successfully.
Create a file named healthCheck in the /opt/vmware/idm/health directory of the vc-ws1a-broker container!
check service status:
service-control --status
cd /storage/containers/vc-ws1a-broker/<container-id>/rootfs/opt/vmware/idm/health
touch healthCheck
This method also works for fresh installations! Because it also stopped here, there is a problem with the 8.0.2 installation package, and they released it without sufficient testing.
Just curious for the folks hitting this, how are you deploying VCSA? Is this a new deployment and if so, what are you using for your networking, specifically DNS as it seems that's where this is getting hung up from one of the responses?
I am experiencing the same issue.
Installing new vCenter Server 8.0 U2 on a new ESXi host with Dell image for 8.0 U2.
Still researching at my end.
Thanks for lamw's reply.
ESXi and vCSA are fresh install and deployment.
Btw, DNS-specific issues additional description: we never deploy DNS server for ESXi and vCSA infrastructure. There's never been any problems with the version (6.5, 6.7, 7.0.3, 8.0.1) we had installed.
thanks for your reply.
VCSA has been successfully installed and run in your way. Thank you!
While you necessary need to deploy a DNS Server to resolve your infrastructure, typically you'll have a DNS resolver for VCSA to resolve external resources like patch infrastructure (which by default will use VMware). So what values are you using when it ask for DNS Server? Are you using a dunmy or non-pingable address?
It's another issue. Regarding the update task, we have always used the offline method.
We don't like to login ESXi or vCSA via DNS resolve.
Creating a DNS host just for ESXi and vCSA is a meaningless task.
We really like simply using IP for ESXi and vCSA management.
In fact, we have used several versions (6.5, 6.7, 7.0.3, 8.0.1) and it's been working for a long time.
Please don't modify this advantage, VMware's engineers.
I have tried both with and without a dns server.
The following is when I have the dns server:
1. I use the dnsmasq that comes with openwrt to provide dhcp and dns services.
2. DNS will automatically associate and bind hostname and IP.
For this I set in the openwrt panel
hostname: vcsa
IP: 192.168.1.15
mac: vcsa mac address
3. I set the local domain and Local server to home.arpa. used to resolve computer names in the LAN, etc.
4. I manually added an A resolution such as vcsa.number201724.com to the vcsa machine, and bound a fixed IP and hostname to the mac address.
5. I manually set vcsa’s fqdn to vcsa.number201724.com and manually set a static ip address.
6. vcsa did not read the fqdn I set, but re-obtained the fqdn from the dns server and changed it to vcsa.home.arpa. (WTFFFFFFFFFFF)
7. I did not check the nginx configuration file carefully. I just saw that the nginx errorlog kept prompting healthCheck and retried until it timed out, causing the installation to fail.
There is no problem with 7.x 8.0 8.0.1, but 8.0.2 always has problems, cannot be installed, and cannot be upgraded.
vcsa:
ip: 192.168.1.15
fqdn: vcsa.number201724.com
C:\Users\number201724>nslookup -q=ptr 192.168.1.15
Server: OpenWRT.home.arpa
Address: 192.168.1.1
15.1.168.192.in-addr.arpa name = vcsa.home.arpa
C:\Users\number201724>nslookup vcsa.number201724.com
Server: OpenWRT.home.arpa
Address: 192.168.1.1
Name: vcsa.number201724.com
Address: 192.168.1.15
C:\Users\number201724>nslookup vcsa.home.arpa
Server: OpenWRT.home.arpa
Address: 192.168.1.1
Name: vcsa.home.arpa
Address: 192.168.1.15
Hi Guys,
I had the same issue here. In my case (lab environment), I was trying to update the VC 7.0.3 to the VC 8.0.2:
From: vCenter Server 7.0 Update 3n
To: vCenter Server 8.0 Update 2
I mounted the VC 8.0 U2 ISO file and ran the "installer.exe" choosing the Upgrade option. During stage 02, I had the same error "An error occurred while starting service 'vc-ws1a-broker'".
I tested 2 solutions to fix this issue:
1. After stage 01, the upgrade process will shut down VC 7.0.3 and "transfer" the FQDN and IP address to the new VC (8.0.2). At this time, I accessed the new VC by SSH and removed the line "search ." on the /etc/resolv.conf file. It worked fine here!
2. The second test that I did was to edit the /etc/resolv.conf file on the VC 7.0.3 before starting the upgrade process. I added the below line on the /etc/resolv.conf:
echo "search lab.local" >> /etc/resolv.conf
Note: lab.local is my local domain name --> This is a lab environment.
I know that is a bad way to edit the /etc/resolv.conf, but for this purpose, it worked fine!
After that, I started the "installer.exe" and did the upgrade process. The upgrade occurred perfectly!
I do not know if there is an issue related to the DNS or some lookup process that depends on the "search" parameter on the /etc/resolv.conf, but these "tips" solved the issue here!
To add the "search" parameter permanently to the /etc/resolv.conf:
-- Making a backup of the file /etc/systemd/resolved.conf:
cp /etc/systemd/resolved.conf /etc/systemd/resolved.conf.backup
-- Editing the file /etc/systemd/resolved.conf and adding the "Domains" instruction, for example:
Domains=lab.local
-- Restarting the "systemd-resolved" service:
service systemd-resolved restart
Option 2 (see below) from DaniloPC verified, upgrade from 7u3 to 8u2 was now successful. Thanks
... Option 2. The second test that I did was to edit the /etc/resolv.conf file on the VC 7.0.3 before starting the upgrade process. I added the below line on the /etc/resolv.conf:
echo "search lab.local" >> /etc/resolv.conf